Two-step knowledge extraction method for knowledge base

• Main Authors: DONG XIAOMING, YING JIE, ZHANG SHU, FAN HUILI, WANG TONG
• Format: Patent
• Language: Chinese; English
• Published: 17.11.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a two-step knowledge extraction method for a knowledge base. The method comprises the following steps: establishing an inverted index table for unstructured, semi-structured and structured data in a ship knowledge base; the inverted index table stores position mapping of words in the document; based on a polling-based load balancing algorithm, searching candidate knowledge of the request problem from the inverted index table cluster; performing a coarse arrangement operation on the searched candidate knowledge; selecting the front candidate knowledge from the roughly arranged candidate knowledge to perform fine arrangement operation, and returning the candidate knowledge screened again; performing triple extraction on the sorted TopN candidate knowledge; performing error correction and knowledge fusion on conflicting triads in the extracted triads; and obtaining a knowledge graph after knowledge extraction according to the processed triple. According to the method, the candidate knowle